You are on page 1of 50

Before the

DEPARTMENT OF COMMERCE
NATIONAL TELECOMMUNICATIONS AND INFORMATION ADMINISTRATION
Washington, DC

In the Matter of

American Recovery and Reinvestment Act Docket No. 090309298-9299-01


Broadband Initiatives

COMMENTS OF THE FIBER-TO-THE-HOME COUNCIL


IN RESPONSE TO REQUEST FOR INFORMATION

Thomas W. Cohen
Kelley Drye & Warren LLP
3050 K Street NW, Suite 400
Washington, DC 20007
(202) 342-8518 (telephone)
(202) 342-8451 (facsimile)
TCohen@a kelleydrye. com_
Counsel to the Fiber-to-the-Home
Council

April 13, 2009


TABLE OF CONTENTS

Page

Summary ......................................................................................................................................... i

1. NTIA Broadband Technology Opportunities Program ...................................................... 2


A. The Broadband Grant Program: A Unique Opportunity to Invest Federal
Funds in Future-Proof Infrastructure ..................................................................... 2
B. The Purposes of the Grant Program: Projects that Further Economic
Growth and Job Creation Take Precedence ........................................................... 3
C. Defining Unserved and Underserved Areas ............................. ............................. 7
D.
14 The Role of the States ....................................................................... ...................
E. Eligible Grant Recipients ..................................................................................... 15
F. Eligible Projects and Use of Grant Funds for Unserved and Underserved
Areas .................................................................................................................... 16

G. Allocating Funding Between Unserved and Underserved Areas ........................ 17


H. Streamline Construction ....................................................................................... 18
1. Establishing Selection Criteria for Grant Awards in Unserved and
Underserved Areas ............................................................................................... 19
J. Non-Discrimination and Interconnection Contractual Conditions ...................... 23
II. Conclusion ....................................................................................................................... 24

APPENDIX
Economic Effects of Tax Incentives for Broadband Infrastructure Deployment, Empiris, LLC

SUMMARY

The Fiber-to-the-Home ("FTTH") Council is a non-profit organization dedicated to


educating the public and government officials about FTTH and to promoting and accelerating
FTTH deployment and the resulting quality of life enhancements FTTH networks make possible.
The FTTH Council's members, including approximately 150 service providers, represent all
areas of the broadband access industry, including telecommunications, computing, networking,
system integration, engineering, and content-provider companies, as well as traditional service
providers, utilities, and municipalities.

At the beginning of the American Recovery and Reinvestment Act of 2009 ("ARRA"),
the "economic stimulus" purposes and principles of the legislation are clearly enunciated, and
these are reiterated in the March 20, 2009 Presidential Memorandum. Therefore the NTIA
should give priority to broadband projects that create the most jobs, deploy infrastructure that
provides long-term economic benefits, and can be initiated promptly by experienced entities.
There is substantial evidence to support the conclusion that deployment of FTTH infrastructure
(in unserved and underserved areas) best fits these ARRA objectives.

The FTTH Council believes that "broadband service" and "unserved" and "underserved"
areas should be defined based on: (1) the economic stimulus objectives of the ARRA; (2) the
policy objective of ensuring there is universal access to broadband services; and (3) the type of
broadband service that is being offered in the market today and during the period when the grants
will be awarded and funding expended. Accordingly, it proposes that unserved areas are those
where a significant number of customers lack access to current generation (6 Mbps/1.5 Mbps)
broadband service and underserved areas include those where a significant number of customers
lack access to a competitive provider of current generation broadband service or any provider of
advanced broadband service (25 Mbps/ 6 Mbps). By using these definitions, grants should be
awarded to projects for deployments of more advanced broadband networks, which should be
beneficial to the creation of jobs and the development of infrastructure with long term benefits.

As with these definitions, the system for awarding grants (the scoring system) should be
based largely on the economic stimulus objectives of the ARRA. The FTTH Council therefore
submits the following system for scoring applications for projects in unserved and underserved
areas: Job creation (up to 20 points); Project feasibility, initiation, and completion (up to 25
points); Infrastructure capabilities and long-term sustainability (up to 30 points); Cost-effective
deployments (up to 10 points); Affordability (10 points); and State endorsements, community
connectivity/support, and disadvantaged small businesses (up to 5 points).

Finally, the NTIA should encourage the submission of projects that are "turn-key" or
involve "design-build" construction techniques by experienced providers. This will enable
projects to be initiated and completed more quickly and will assist the agency in monitoring
compliance.

Before the
DEPARTMENT OF COMMERCE
NATIONAL TELECOMMUNICATIONS AND INFORMATION ADMINISTRATION
Washington, DC

In the Matter of

American Recovery and Reinvestment Act Docket No. 090309298-9299-01


Broadband Initiatives

COMMENTS OF THE FIBER -TO-THE-HOME COUNCIL


IN RESPONSE TO REQUEST FOR INFORMATION

The Fiber-to-the-Home Council ("FTTH Council"), through its undersigned

counsel, hereby respectfully submits its comments to the Department of Commerce,

National Telecommunications and Information Administration ("NTIA") in response to

the March 12, 2009 Federal Register notice for comments ("Notice") to implement the

Broadband Technology Opportunities Program (`STOP") in the American Recovery and

Reinvestment Act of 2009 ("ARRA").'

The FTTH Council is a non-profit organization established in 2001. Its mission is

to educate the public and government officials about fiber-to-the-home ("FTTH") and to

promote and accelerate FTTH deployment and the resulting quality of life enhancements

FTTH networks make possible. The FTTH Council 's members represent all areas of the

broadband access industry, including telecommunications, computing, networking,

1 In the Matter of the American Recovery and Reinvestment Act of 2009 Broadband
Initiatives, Request for Information, Docket No. 090309298-9299-01, Rel. March
12, 2009. The BTOP is established in Section 6001 of the ARRA.

system integration, engineering, and content-provider companies, as well as traditional

service providers, utilities, and municipalities.'

I. NTIA Broadband Technology Opportunities Program

A. The Broadband Grant Program: A Unique Opportunity to Invest


Federal Funds in Future-Proof Infrastructure

For years, the United States has been slipping behind other countries in deploying

broadband networks and services. The Information Technology & Innovation

Foundation in 2008 ranked the United States 15th, lagging many Asian and European

countries.3 The Organization for Economic Co-operation and Development reached a

similar conclusion,4 and the just issued report from Akamai found that the United States

ranked 17th in average internet access speed.5

More importantly, other countries are not standing still. Instead, they recognize

that advanced broadband infrastructure is fundamental to their ability to foster business

development and economic growth and are accelerating their efforts to build these

networks. Australia, for instance, just announced a program to spend $31 billion to bring

FTTH - with internet access at 100 megabits per second -- to 90% of the nation's

z
As of today, the FTTH Council has more than 200 entities as members. A
complete list of FTTH Council members can be found on the organization's
website, httm//NN-,v w.ftthcounciLp.Eg.
3
See,httpJ;',,vww.i.tif.org/filest2008BBR.ankings.pf. The ITIF rankings are based
on penetration, speed, and price.
4
See, http:/ ^vww.itif.org/:files/BroadbandRaiikings.pdf.
5
See,
http://xk,-,N,w.akamal.com.ihtml/about/pressl'releases/2009,'press 033009 1.html.

households and workplaces in the next eight years. This FTTH project will support up to

25,000 jobs on average for each year of the deployment.6

Fortunately, the federal government has begun to react to our flagging position

with the ARRA and the new broadband programs. We have an opportunity to begin our

effort to keep pace with other countries and achieve the President's objective of

promoting "next-generation [broadband] facilities"7 by spending these funds on "future-

proof' infrastructure, such as FTTH, which can be upgraded readily without major new

construction and which provides long-term benefits for residents, businesses, and our

economy as a whole. This should be the vision for the NTIA as it implements the BTOP.

B. The Purposes of the Grant Program: Projects that Further Economic


Growth and Job Creation Take Precedence.

At the beginning of the ARRA, the purposes and principles of the legislation are

clearly enunciated:

• "Preserve and create jobs and promote economic recovery."

• "Invest in... infrastructure that will provide long-term economic benefits."

• "[Commence] expenditures and activities as quickly as possible consistent

with prudent management."8

These purposes - "stimulating economic growth and job creation" -- are later echoed in

the Conference Report on the provision establishing the BTOP.9 Moreover, the March

6
See, http://www.pm.gov.au/.in.ediaiRelease/'2009,'media release 090 3.cfm.
7
See, http://www.barackobama.com/pdf/issues/FactShectScience.pdf.
s
Public Law 111-5, Section 3 (a)(1), (a)(4), (b).
See, Summary, Division B, Title V1, Broadband Technologies Opportunities
Program, Conference Report on H.R. 1, American Recovery and Reinvestment
Act of 2009, February 12, 2009. ("Conference Report")

20, 2009 Memorandum from the President, Ensuring Responsible Spending of Recovery

Act Funds, states that "merit-based selection criteria... shall be formulated to ensure that

the funding furthers the job creation, economic recovery, and other purposes of the

Recovery Act."10 The NTIA accordingly should give priority to broadband projects that:

• Create the most jobs;

• Can be initiated promptly by an experienced entity; and

• Deploy infrastructure that spurs economic development.

There is substantial evidence to support the conclusion that deployment of FTTH

infrastructure (in unserved and underserved areas) best fits these objectives. First, in

terms of immediate jobs and economic output, FTTH deployments are enormous

construction projects, involving far more outside plant work than other technologies. I I

This conclusion is supported by a recent study by the economic consulting firm Empiris,

LLC, which was commissioned by the FTTH Council, on the economic effects of tax

incentives for the deployment of broadband infrastructure. (A full copy of the report is

appended to these comments.) Of particular relevance to the discussion here, the report

states:

[A] majority (54 percent) of capital spending required in outside plant


build-out for FTTH is spent on construction. This heavy reliance on

Io
Memorandum for the Heads of Executive Departments and Agencies, Subject:
Ensuring Responsible Spending of Recovery Act Funds, The White House,
released March 20, 2009, Sec. 1.
II
This conclusion is supported in the Grant Distribution Considerations and
Broadband Speed, Title VI section of the Conference Report, which provides,
"The Conferees are also mindful that the construction of broadband facilities
capable of delivering next-generation broadband speeds is likely to result in
greater job creation and job preservation than projects centered on current-
generation broadband speeds."

4
construction for FTTH is due in large part to the burying of new
infrastructure in the ground...$1 million of investment in FTTH
deployment will result in almost 20 jobs, whereas a million dollars of
investment in wireless broadband will result in fewer than 15 jobs. This is
largely due to our estimate that only 7 percent of wireless broadband
capital expenditures go to the construction industry.

In addition to the immediate benefits of creating economic growth and jobs, in terms of

economic development, FTTH deployments provide by far the most capabilities (through

higher symmetrical bandwidth) for customers to send and receive data and video, and

these networks are "future-proof." Once the fiber is installed, upgrading the capabilities

of the network is readily accomplished by changing the electronics. In addition, fiber

networks are most valuable to businesses, which increasingly demand dedicated amounts

of bandwidth, and to their employees.12 As the former Mayor of Ft. Wayne, Indiana,

Graham Richard, stated on March 16, 2009, at a meeting of the National League of

Cities, "If you don't have [FTTH], [companies] won't invest in your city. [Broadband

deployment] is just as important as public safety, water and sewer systems."13

Because of the paramount importance of the ARRA's economic stimulus

objectives, in determining which of the five purposes set forth in Section 6001(b) of the

BTOP, the NTIA should favor - and expend more funding on - those purposes that best

achieve these overriding objectives. If there are too many projects that achieve all of

12
A 2007 survey sponsored by the FTTH Council found that "a substantial portion
of Americans who get their home Internet services through direct fiber optic
connections are using those services to telecommute an average of one-third of
the time or to run their own home-based businesses," See,
http:// ww.ftthcouncil.org;'?t=262.
13
See, TR Daily, Local Involvement Urged in ARRA Grant Program, March 16,
2009. It should be noted that Mayor Richard stated that job creation was the
"number one" benefit from FTTH deployment.

5
these objectives and the NTIA needs to examine other, secondary factors, it is important

to note that the statutory text of Section 6001 does not favor one purpose over another.

Thus, once the overall purposes of the ARRA are met, the BTOP provisions do not

dictate that one purpose should outweigh another. The Conference Report sheds some

light on how to proceed by stating that the "NTIA has discretion in selecting the grant

recipients that will best achieve the broad objectives of the program." 14 In sum, the

NTIA is not obligated to favor one purpose over another and has significant discretion in

determining how to allocate the funding -- with one critical exception, it must first favor

those projects that meet the primary "economic jobs" purposes of the ARRA.15

The Conference Report provides insight as to the intentions of the Conferees as to

which projects best achieve these primary purposes - "[T]he construction of broadband

facilities capable of delivering next-generation broadband speeds is likely to result in

greater job creation and job preservation than projects centered on current-generation

broadband speeds. Therefore, the Conferees instruct the NTIA to seek to fund, to the

extent practicable, projects that provide the highest possible, next-generation broadband

speeds to consumers." 16 As discussed above, the report by Empiris, LLC provides

support for this conclusion that FTTH deployments generate the most jobs and economic

14
Summary, Title VI, Conference Report.
15
More specifically, there is no indication in the statute or Conference Report that
unserved areas should be favored over underserved areas in awarding grants, or
vice versa - nor does the legislation provide any criteria for NTIA to use in
establishing priorities. In fact, the Conference Report (Summary, Title VI) states
that the "Conferees intend that the NTIA award grants serving all parts of the
country, including rural, suburban, and urban areas."
16
Grant Distribution Considerations and Broadband Speeds, Title VI, Conference
Report.

output. Further, FTTH networks are critical for subsequent economic development

because of their great capabilities for businesses and their employees, community and

public safety organizations, educational institutions, faculty, and students, and heath care

institutions, their personnel, and patients.

C. Defining Unserved and Underserved Areas

In the Notice, the NTIA asks for comments on defining two critical terms:

unserved and underserved areas. Neither the statute nor the Conference Report provides

any insight into how these terms are to be defined, and the report merely instructs the

NTIA to consult with the FCC. Of these two terms, an unnerved area is somewhat easier

to define - that is, the natural reading of the term is that it is an area without any

broadband service. Two questions then arise: what is broadband service in this context,

and what number of customers in an area need to receive that service for the area to be

declared served? The FTTH Council submits that, for unserved areas, the aim of the

BTOP should be to provide on a universal basis at least the same type of broadband

service most generally available in served areas. For wireline and other fixed broadband

service, there is sufficient evidence to conclude that the "current" generation broadband

service that is generally available in served areas has at least the following performance

characteristics: 6 Mbps (Rate Code 6) downstream and 1.5 Mbps upstream (Rate Code

4).17 (The FTTH Council believes the information collected by the FCC in its FCC Form

17
The FTTH Council submits that providing an objective measure of broadband
performance is essential to the success of the BTOP. It will enable the agency to
judge applications on their merits, and, just as importantly, ensure that entities
receiving grants are properly monitored. While determining performance by
broadband speed alone may not be optimal, there is no other objective standard.
Moreover, it is the standard by which current broadband service providers
... Continued

477 will be relevant to the NTIA as it reviews applications and therefore links its

recommended performance characteristics with the FCC's Rate Codes.) There is

considerable evidence to support this conclusion. For instance, a recently issued report

from the consulting firm, Point Topic, found that in the 4t" quarter 2008 the average cable

download speed was 9.7 Mbps, while the digital subscriber line ("dsl") speed was 3.9

Mbps.18 This is supported by examining the offerings of individual providers. Cox's

Preferred service has speeds of 10 Mbps downstream and 2 Mbps upstream. 19 Verizon,

for its dsl service, routinely offers between 3-7 Mbps downstream and approximately 1

Mbps upstream. 20 It is thus not surprising that a just published New York Times article

reported, "The United States has an average [downstream] speed of 5.2 Mbps."21 Finally,

it should be noted that similar speeds (5 Mbps/1 Mbps) were included in the House

passed ARRA legislation (H.R. 629) in its definition of basic broadband service. 22

As for the second issue of the "sufficiency of current generation service" - that is

determining how many customers in an area need to lack access to broadband service for

the area to be declared unserved -- the FTTH Council submits that the objective of the

Program should be to achieve "universal" broadband service. One methodology to define

advertise and offer their services - and it is the standard used in the Conference
Report.
18
Data supplied by Point Topid Ltd. See, http:,•''/pornt-topic.com/.
19
See, h.ttp:,'/Arw2.cox.com/ €•esidential; connccticut'i.nter.net/preferred-inter€Zet.cox.
20
Verizon offers a Power Plan (3 Mbps/768 Mbps) and a Turbo Plan (7.1 Mbps/
768 Mbps) See,
httn_'/www22.verizon.com/Residenti.aL `Hi hS eedlnternet'Pl.an.s'Plans.htm.
21
See, http://bits.blogs.nytimes.conif2009/'03/`10/tlie-broadband-gap-why-is-theirs-
fasterr'.
22
Sec. 10020)(3).
"universal" is to examine the current nationwide reach of broadband service and use that

as a benchmark. In that regard, the Federal Communications Commission, in a January

2009 report, found that people in 9% of the zip codes in the United States do not have

access to either cable modem or ds1.23 This statistic is supported by the cable industry,

which just stated that cable operators reach over 90% of the households in the United

States.24 Thus, there is sufficient evidence to conclude that a Census Tract should be

declared unserved where more than 10% of the customers lack access to broadband

service. The FTTH Council has proposed in its rules and comments submitted for the

Rural Utilities Service program a slightly more conservative approach -- a standard of

"20% lacking access" -- but it does not object to the more rigorous approach to ensuring

universal service. This will ensure that broadband service is brought to almost all users

in the country.

Unserved Area means a geographic area described by Census Tracts where more
than 20% of the customers (either residential or business or both) to be served by
the project currently lack access to a provider of Broadband Service. 25

Defining underserved areas is more challenging. As noted above, neither the

statute nor the Conference Report nor the original House and Senate bills provides any

indicia of Congressional intent. Some context for the meaning of the term can be gleaned

from the purpose of the statute - "provide improved access to broadband service" - and

2323
High Speed Services for Internet Access: Status as of December, 31, 2007,
Industry Analysis and Technology Division, Wireline Competition Bureau,
Federal Communications Commission, January 2009 at Table 16.
24
See, Moving the Needle on Broadband at
http:/,',^vww.:n.eta.com/PublicationTypc /W.hite.Paper.,"Movirlg-the-Needle-on-
Broadband.aspx.
25
Either current generation or advanced broadband service.

from statutory criteria for awarding grants - "provide the greatest possible speed" and

"increased affordability. "26 There is precedent for using in the definition both the

concept that an underserved area is one where access to advanced broadband service is

insufficient and where there are select groups of customers needing more affordable

access to broadband service. For instance, the concept of insufficient access to advanced

broadband service is used in the definition of underserved adopted for the California

Advanced Services Fund grant program and in broadband tax legislation introduced this

year in the U.S. House of Representatives. 27 Other broadband tax legislation introduced

this year in the House and Senate keys the term to groups of customers where

affordability is a concern. 28

Because the overall purpose of the ARRA is to stimulate the economy and

provide long-term infrastructure, the FTTH Council encourages the Agency to adopt a

definition for underserved areas that would enable projects that achieve those essential

economic stimulus objectives - and that melds them with the two concepts of bringing

sufficient access to advanced broadband service to areas and in ensuring affordable

access by select groups of customers. Accordingly, the FTTH Council believes the

definition of underserved area should be the following:

26
Sec. 6001(b)(2) and (h)(2)(A) and (B). The FTTH Council believes that any
affordability standard needs to be judged on the basis of the performance of the
broadband service required, and, in the application scoring system discussed later,
uses the concept of cost per service as represented by performance speed (megabit
per second).
27
Resolution T-17143. Approval of the California Advanced Services Fund (CASF)
Application Requirements and Scoring Criteria forAwarding CASF Awards,
Public Utilities Commission of the State of California, June 12, 2008, p.6. H.R.
760, Advanced Broadband Infrastructure Bond Initiative of 2009, 11 It" Congress.
28
See, S. 350, Sec. 48D(e)(24), and H.R. 691, Sec. Sec. 45R(c)(3).

Underserved Area means:


(1) a geographic area described by Census Tracts that is not an unserved area
where more than 33% of the customers (either residential or business or both) to
be served by the project currently lack access to more than one provider of
Current Generation Broadband Wireline Service;
(2) a geographic area described by Census Tracts that is not an unserved area
where more than 33% of the customers (either residential or business or both) to
be served by the project currently lack access to a rovider of Advanced
Broadband Wireline Service;
(3) a geographic area described by Census Tracts where more than 25% of the
Community Anchor Institutions to be served by the project currently lack access
to a provider of Advanced Broadband Wireline Service; or
(4) any Census Tract that is located in (1) an empowerment zone or enterprise
community designated under section 1391, (ii) the District of Columbia
Enterprise Zone established under section 1400, (iii) a renewal community
designated under section 1400E, or (iv) a low-income community designated
under section 45D.29

The first part of the proposed definition - (1) - seeks to ensure affordability by

encouraging greater competition for an entire area. The other parts - (2), (3), and (4) --

of the proposed definition are directed towards ensuring that most customers in general

and key segments of the community in particular have "first-time" access to advanced

broadband service. This part not only achieves a universal service objective but is most

important in meeting the ARRA's economic objectives since it will foster the deployment

of new advanced broadband networks, which as discussed above in the case of FTTH

networks are significant generators of jobs and economic growth - as well as the

construction of infrastructure with long-term benefits.

There are two terms used in the proposed definition that require elaboration - the

definition of advanced broadband service and the issue of insufficient access to such
service. First, as to the definition of Advanced Broadband Wireline and other Fixed

Service, the FTTH Council proposes the following definition:

Advanced Broadband Wireline and other Fixed Service means providing on an


advertised and generally available basis a dedicated service to each customer from
the internet access node an information transfer rate equivalent to at least 25
megabits/second from the provider to the customer (downstream) and at least 6
megabits/second from the customer to the provider (upstream).

This definition is based on current performance characteristics for high-speed broadband

services many customers receive or about to receive. (The downstream speed fits within

Rate Code 8 of the FCC Form 477 and the upstream speed within Rate Code 6.)

Verizon's FiOS network, for instance, today provides a 50 Mbps downstream - 20 Mbps

upstream service, 30 and AT&T's U-verse today provides a 18 Mbps/1.5 Mbps service. 31

In addition,32 Comcast offers a 50 Mbps/10 Mbps service;33 Cablevision a 30 Mbps/5

Mbps service, 34 and Cox just launched a 50 Mbps/5 Mbps service. 35 These performance

29
For both unserved and underserved areas, the Agency should ensure that any
existing provider is viable and that its service is sufficiently substitutable with the
broadband service proposed in the application.
30
See, http__1'i ivestor_ve_rizon..com/news/vicw.aspx?Newsl:I^_t)25.
31
See, ..http:;;h,,-w w.att-services..n.et/att-u-verse/uverse-internet.11tml.
32
It should be noted that the performance metrics offered by cable operators are
based on providing a shared service, that is a service where all customers do not
receive these high-speeds simultaneously. In its proposed definition of advanced
broadband wireline service, the FTTH Council provides that the 25 Mbps/6 Mbps
performance requirements must be met by providing a dedicated service to ensure
all customers receive the proposed speeds simultaneously.
33
See, http://www.cmcsk.com'phoenix.zhtml?c=118591 &p=irol-
newsA.rtiel.e&ID=1215838&higLhligh_t=. Comcast's current, general high-speed
offering is 16 Mbps/2 Mbps.
34
See, http:!:-www.Optil-n:um.com/order/boost/.
35
See, l:lttp://sev.pmews\ ire.com/'comptrter-
e:lectrotlies/20090401 iCl:.,0228901.042009-1..html..

characteristics are, of course, snapshots of the current market, and, given the previous

growth rates, the Agency should conclude that even higher-speeds will be available in the

next few years - when the networks built by the broadband stimulus grants are

completed. There is thus sufficient evidence for the Agency and the Commission to

conclude that advanced broadband service should be defined as at the least as 25 Mbps

downstream - 6 Mbps upstream.

Additional support for this definition, reflecting recent Congressional intent, can

be found in the definition for "advanced broadband service" used in the House version of

the NTIA broadband stimulus program in the AR.RA - "at least 45 megabits per second

downstream and at least 15 megabits per second upstream."36 While this definition was

not included in the final bill, Congress required the NTIA to favor applications for grants

that offered the "greatest possible broadband speeds. ,37 Clearly, Congress understands

that higher speeds provide the greatest capabilities for customers - in addition to leading

to the greatest benefits in terms of economic growth and job creation. The FTTH Council

would welcome the Agency's adoption of the definition provided by the House but also

believes the definition provided above is sufficient to achieve the ARRA's intent.

The other term that requires elaboration is the threshold - 33% of the customers

lacking access -- for triggering an area to be underserved. This determination should be

driven by the overall objective of bringing to underserved areas the same level of service

- that is, advanced broadband service -- found in served areas. For advanced broadband

service, the major broadband providers will pass a majority of their customers in the next

36
See, Section 10020), H.R. 629, 11 1`h Congress.

year, and this level of access will continue to grow. In terms of a current snapshot,

Verizon's FiOS network will pass approximately 18 million households by the end of

2010 - over 50% of the households it serves. 38 Comcast, the nation's largest cable

company, advanced broadband service will be provided to approximately 10 million

homes shortly. 39 Cablevision, another major cable operator, currently offers most of its

customers its high-speed service.40 AT&T's U-verse will pass approximately 55% of the

households it serves by the end of 2010.41 It is on this basis - of generally available

advanced broadband service offered during the period when infrastructure using grant

funding from the Program will be built -- that the FTTH Council submits that the area

should be declared underserved if more than 33% of the customers lack access to

advanced broadband service.

D. The Role of the States.

The statute and the Conference Report clearly provide that authority and

responsibility for awarding grants under the BTOP rests solely with the NTIA and that

37
See, Grant Distribution Considerations and Broadband Speeds, Title VI,
Conference Report.
38
See, httl?Jiin.vestol•.vei-izon.cominews/view.aspx?_NNewsID=925.
39
See, httT):://ww cmcsk.comlphoenix.zhtml?c=_118591 &p= irol-
newsAltic(e&ID __121_5838&highli.t=. An article in Communications Daily on
April 13, 2009 (p. 6) states that Comcast "is aiming to wire 65 percent of its
footprint for wideband by the end of this year, and expects to complete the
deployment by sometime next year."
40
See, http://www.optii-rium.com/order/boost/`.
41
See, llttp://www.att-services.net/att-u-verse/uverse-intemet..html and
http://www.att.com/gen/investor-relations?pid=571 1.
the Agency has discretion in considering any state submissions.42 States may identify

and supply to the NTIA information about unserved and underserved areas. States also

may provide advice about how to allocate funding among projects in the state -but,

again, the NTIA has the ultimate responsibility (and discretion) and must adhere to the

objectives of the ARRA. Thus, should the NTIA decide to consult with a state, for a

state's submission to carry any weight, it needs to be accompanied by supporting

documentation that sufficiently sets forth the bases for its conclusions and, where

relevant, how those conclusions meet the purposes of the ARRA. Those submissions that

meet these requirements should be accounted for in scoring applications.

The statute also provides that states may apply directly for grants. When this

occurs, state filings should be treated the same as any other filings and should be judged

solely on the merits of the project submitted.

E. Eligible Grant Recipients.

The statute gives the NTIA the discretion to expand the eligible grant recipients to

any entity so long as that determination is in the public interest and to the extent

practicable promotes the purposes of the BTOP in a technologically neutral manner. The

Conference Report provides further insight by stating that the Conferees intend "that,

consistent with the public interest and purposes of this section, as many entities as

possible be eligible to apply for a competitive grant." To ensure that the projects that

42
The statutory text states that NTIA "may consult a State," implying the agency
has significant discretion (Sec. 6001(c)). The Conference Report appears to
support this approach, although the language is more opaque since it states that
the conferees "expect and intend that the NTIA, at its discretion, will seek advice
and assistance from the States." In any event, since the statutory language is
clear, there is no need to refer to the report.

best achieve the purposes of the statute are submitted, the FTTH Council supports

opening the grant process to as many legally-organized entities as possible.

F. Eligible Projects and Use of Grant Funds for Unserved and


Underserved Areas.

The eligibility of projects in unserved and underserved areas should be based on

the objectives of the BTOP and flow directly from the definitions of those areas.

Consequently, in unserved areas, projects should seek - and funds should be made

available - to provide either Current Generation or Advanced Broadband Wireline

Service, although grantees that seek to deploy advanced service should be preferred. In

underserved areas, project funding should only be for the provision of Advanced

Broadband Wireline Service. In addition to these requirements, the Agency should

permit any project to serve customers in unserved and underserved areas to include so-

called high-speed middle-mile connections, that is from the access plant to the internet

node. Middle-mile facilities, which may be costly to construct especially in rural areas,

are critical to ensuring customers do not face "broadband bottlenecks." Moreover, the

effectiveness of other demand-side and infrastructure projects funded under BTOP could

be severely diminished without adequate investment in middle-mile facilities. Finally,

funding should be directed towards infrastructure deployment with the exception of

assisting in ensuring access to community anchor institutions and vulnerable populations.

The proposed FTTH Council rule is as follows:

(a) ELIGIBLE GRANT PURPOSES FOR PROJECTS IN UNSERVED AND


UNDERSERVED AREAS
Grant funds may be used to finance:
(1) In unserved areas, or in underserved areas (defined by enabling a
second entrant), the acquisition of equipment, instrumentation,
networking capability, hardware and software, digital network
technology, and infrastructure used in the provision of Current

Generation or Advanced Broadband Wireline Service and the


construction and deployment of such service related infrastructure;
(2) In underserved areas (where no or insufficient Advanced
Broadband Wireline Service), the acquisition of equipment,
instrumentation, networking capability, hardware and software,
digital network technology, and infrastructure used in the provision
of Advanced Broadband Wireline Service and the construction and
deployment of such service related infrastructure;
(3) To ensure the service performance requirements in (1) and (2)
above are met, the acquisition of equipment, instrumentation,
networking capability, hardware and software, digital network
technology, and infrastructure and the construction and
deployment of such service related infrastructure to connect access
networks with the point of access to the internet (middle-mile
connections);
(4) Access to Broadband Service, including use of end-user
equipment, by Community Anchor Institutions;
(5) Access to Broadband Service, including use of end-user
equipment, by low-income, unemployed, aged, and otherwise
vulnerable populations; and
(6) The purchase of land, buildings, or building construction needed to
carry out the project.
(b) INELIGIBLE GRANT PURPOSES
For projects seeking to provide Broadband Service in unserved or
underserved areas, operating expenses incurred in providing such service
are ineligible for grants, except as provided in subsection (a)(3) and (a)(4)
of this section.

G. Allocating Funding Between Unserved and Underserved Areas.

As for the allocation of funding between unnerved and underserved areas, under

the statute, the NTIA has discretion but must act consistently with the overall intent of the

ARRA, that is, it should determine which projects best stimulate the economy.43 In

interpreting this requirement, the Agency first must determine how grant funding will be

expended. For projects in underserved areas (at least for the first two parts of the FTTH

Council's proposed definition), entities will need to deploy infrastructure over which the

43 The FTTH Council suggests for the allocation of the overall funding that the
agency seek to award most of the money in the first two of the three rounds of
funding. This would to maximize the immediate impact on the economy.

provider offers advanced broadband services, which it defines as a service with speeds of

at least 25 Mbps downstream and 6 Mbps upstream.44 As discussed earlier in these

comments, such infrastructure inherently creates more jobs and economic growth than

projects for current generation services - thus better achieving the purposes of the ARRA

- and, as such, should receive a greater allocation of funding. The FTTH Council

accordingly suggests for the initial round that funding the Agency allocate more funds to

underserved areas than to unnerved areas.45 This ratio can be adjusted in subsequent

rounds to ensure the objectives of the statute are best achieved.

H. Streamline Construction

As noted frequently throughout these comments, the objective of the ARRA is to

stimulate the economy by expending funds on worthwhile projects as quickly as possible.

The Agency thus should seek to ensure that projects begin and proceed expeditiously

once grants are awarded. To that end, it should permit entities to file "turn-key" or

"design-build" projects, whereby the provider joins with or otherwise obtains

commitments from equipments vendors, construction companies, and other participants

in the project in advance of filing the application. In essence, all of these entities become

44
As noted earlier in the comments, the House legislation defined advanced
broadband service as 45Mbps/15Mbps. While the FTTH Council supports this
definition because such service is a current offering by many providers over
FTTH networks, it acknowledges that infrastructure capable of providing
broadband service at a speed of 25Mbps/6Mbps also will meet the ARRA's
intent.
45
As noted in its discussion of which entities should be eligible to apply for grants,
the FTTH Council urged an expansive approach which was technology neutral.
The statute, however, does not require that grants be awarded on a technology
neutral basis. Rather, NTIA is to award grants based on the purposes of and
specific requirements in the statute, which, when applied, may result in one
technology being favored over another.

part of the project team and are included in the application as participants. This will

obviate the need for post-award bidding and negotiations between the applicant and

suppliers, which will only delay initiation of the project.

1. Establishing Selection Criteria for Grant Awards in Unserved and


Underserved Areas.

The selection criteria for grant awards needs to flow from the purposes of the

ARRA and the text of the statute. Thus, the major criteria must be: job creation,

economic stimulus, and deployment of infrastructure with long-term benefits. In a

scoring system, these components (and their surrogates) should comprise the majority of

points. The BTOP statute also lists a series of considerations that the Agency must

consider, some of which may overlap with the criteria described above, including:

increase affordability and subscribership; provide the greatest broadband speed to the

greatest population; enhance service for health care delivery, education, or children; and

not result in unjust enrichment as a result of support for non-recurring costs through

another federal program. The FTTH Council appreciates the challenge that the NTIA has

in weighing all these factors, and it proposes the following criteria and points for

evaluating infrastructure deployment projects in unserved and underserved areas (out of a

maximum of 100).46

UNSERVED AND UNDERSERVED INFRASTRUCTURE SCORING


CRITERIA
The following formulas shall be used to score an application that seeks to provide
access in unnerved areas or improved access in underserved areas. Unless
specifically stated otherwise, the same formulas will be used for each services

46
Outside of the context of infrastructure projects dedicated to unserved and
underserved areas, the FTTH Council supports an approach that would
independently address projects focused solely or largely on the two other
important purposes in the statute: education/awareness/training and public safety.

category and for underserved and unserved areas. All eligible applications shall
receive points (Maximum 100 points) pursuant to the following scoring criteria:

SCORING CRITERIA DEFINED


1. Jobs Creation (up to 20 points). The number of jobs directly
involved in the project, including for construction, support, and
management. The points awarded shall be determined by the
following formula:

Jobs Points = (j 1 / tcl) / (Max(j / tc)) * 20

j = Direct jobs proposed to be created by the project


(expressed in person-hours of work to be expended by all
individuals directly working on the project)
tc = Total cost of project
(1) "jl / tcl" is the ratio "j / tc" for the specific project
being scored;
(2) "Max(j / tc)" is the highest ratio "j / tc" of any
project submitted by an applicant during the current
filing window.

2. Project Feasibility, Initiation, and Completion (up to 25 points).


The Agency shall not consider whether the applicant has requested
or receives a waiver of the Matching Funds requirement in making
any of these scoring determinations. The Agency shall not
preclude or otherwise discriminate against applications where
project design, construction, and installation are engaged in
simultaneously or in advance of filing the application, that is,
"design-build" or "turn-key" projects.
A. Feasibility (up to 15 points). This criterion measures the
project's overall chances for successful completion.
Among the factors to be considered are:
(1) The experience of the applicant;
(2) The ability of the applicant to obtain the necessary
labor and materials at the price specified in the
application;
(3) Use of proven technologies; and
(4) Existence of legal barriers.
B. Initiation and Completion (up to 10 points). This criterion
measures the number of months the applicant has proposed
to take to complete the project in light of the total cost of
constructing the project (that is, exclusive of the cost of
materials, including electronics). (Using only the
construction budget will better correlate with the direct jobs
and immediate economic stimulus caused by the project.)
Points shall be determined based on the following formula:

Timeliness Points = (Min (m / tc) / (mI / tc,)) * 10

(1) Where "m" is the number of months between the


date of the grant award and the proposed date of
completion of the specific project being scored;
"ml" is the "m" for the specific project.
(2) Where "tc" is the total cost of construction of the
proposed project; "tcI" is the "tc" for the specific
project. ("tc" should be expressed in millions of
dollars.)
(3) Where Min (m / tc) is the smallest ratio for any
project submitted of the proposed number of
months to complete the project divided by the total
cost. (The ratio should be expressed in months per
millions of dollars.)
(4) The Agency has the discretion to alter "in" for
projects in areas where construction is infeasible
because of inherent weather conditions, e.g. states
in very cold climates. In such instances, the
Agency may alter "m" by delaying the initiation
date of the project from the date of the grant award
to the date when weather permits construction to be
undertaken and may "stop the clock" if the project
is to continue through subsequent periods when
construction is infeasible.

3. Infrastructure Capabilities and Long-Term Sustainability (up to 30


points)
A. Broadband Transmission Speed (up to 20 points). This
criterion represents the difference between the current
average advertised and generally available dedicated
broadband transmission speed per customer and the
applicant's proposed advertised and generally available
dedicated speed per customer in the proposed service areas.
Points will be determined separately for each service
category based on the following formula:
Speed Points = b / Max(b) * 20
(1) "b" is the "b" value for the specific project being
scored;
(2) "b" _ (proposed service upload speed - current
service upload speed) + (proposed service
download speed - current service download speed).
If no current broadband service, the speed shall be
zero. (All speeds used in the formula should be for
the tiers with the maximum speeds which are
advertised and generally available).

(3) Max(b) is the highest "b" value of any project


submitted by an applicant during the current filing
window - to a limit of 50 Mbps downstream and 20
Mbps upstream;
B. Long-Term Sustainability (up to 10 points). The Agency
shall award up to 10 points based on its assessment of the
proposed project's likelihood of sustainability, including
whether the project includes infrastructure that can be
readily upgraded to provide greater performance.

4. Cost-Effective Deployments (up to 10 points). This criteria shall be


determined based on the number of customers the applicant will be
able to serve divided by the funding amount requested from the
Agency. It is conditioned on the requirement that applicants must
provide at least Current Generation Broadband Service in
Unserved Areas, and Advanced Broadband Service in Underserved
Areas. (The Agency can modify this requirement to account for
the provision of speeds in excess of the required speeds by
including an additional multiplier similar to the one included in the
3(A) - Broadband Transmission Speed (b/Max(b))). Points will be
determined based on the following formula:

Unserved Areas:

FRPPC Pointstillserved =
((Min(a) / a,) * (Min(d) / di)) * 10

Underserved:

FRPPC Pointsunderserved =
(Min(a) / a,) * 10

a= Funds Requested per Potential Customers (FRPPC)


expressed as dollars per household passed; "d" is
the average number of homes per square mile in the
proposed service territory.
(1) "a, " is the "a" for the specific project being scored;
(2) "Min(a)" is the lowest "a" value of any project
submitted by an applicant during the current filing
window;
(3) A density factor is included for unserved areas,
which are most likely to be rural, to account for
higher cost in these areas. "d," is the "d" for the
specific project being scored;
(4) "Min(d)" is the lowest "d" for any project submitted
by an applicant during the current filing window.

5. Affordability (10 points). 10 points shall be awarded to an


applicant that commits to charge each customer served over the
infrastructure constructed as part of a grant a price /Mbps for
Broadband Service equal to or less than the average price /Mbps for
existing internet access service (whether broadband or dial-up
service) in the area covered by the project for a period of at least
two (2) years after service is initiated. (All prices and speeds
(Mbps) should be for the tier with the maximum speed which is
advertised and generally available.)

6. State endorsements, community institution connectivity and


support, and socially and economically disadvantaged small
business concern (up to 5 points). In the discretion of the Agency,
applicants may receive up to 5 points for obtaining support from a
Governor of a state (or the Governor's designee), connecting
community anchor institutions, obtaining community support, or
demonstrating that it is a socially and economically disadvantaged
small business concern.

J. Non-Discrimination and Interconnection Contractual Conditions.

The statute requires the NTIA publish non-discrimination and network

interconnection obligations, which at a minimum shall include the FCC's Wireline

Broadband principles.47 These obligations will be contractual conditions of grants. The

Conference Report provides no additional commentary on the Conferees intent on this

issue. The FTTH Council urges the NTIA to act conservatively in implementing this

requirement and limit the obligation to the FCC's principles for two primary reasons.

First, the main purpose of the ARRA is to stimulate the economy, and the creation of a

new set of obligations will take additional time, generate uncertainty, and potentially

foster disputes - all of which will slow implementation. The FCC's principles, in

contrast, are a known quantity. Second, there is a consensus in the public and private

sectors in support of the FCC's principles, and there is every indication that providers,

47 FCC Docket No. 05-15, adopted August 5, 2005.


with rare exception, live by them. The number of complaints brought for non-

compliance have been minimal. In addition, there have been few complaints about

practices of providers that may fall outside the principles. In sum, for this targeted

program, the FTTH Council believes relying on current policies has real value.

II. Conclusion

The BTOP provides the Agency with a great opportunity to achieve the economic

stimulus aims of the ARRA and propel the deployment of broadband infrastructure that

provide long-term benefits to residents and businesses - which in turn will enhance our

international competitiveness. The FTTH Council believes that deployments of FTTH

networks best meet these objectives, and it stands ready to assist the Agency as it moves

forward in implementing this Program.

Respectfully submitted,

acv.
Thomas W. Cohen
KELLEY DRYE & WARREN LLP
3050 K Street NW, Suite 400
Washington, D.C. 20007
(202) 342-8518 (telephone)
(202) 342-8451 (facsimile)
TCohe^i)k.elleydrye.com
L

April 13, 2009


ECONOMIC EFFECTS OF TAX INCENTIVES FOR BROADBAND


INFRASTRUCTURE DEPLOYMENT

JEFFREY A. EISENACH
HAL J. SINGER
JEFFREY D. WEST

EMPIRIS, LLC

PREPARED ON BEHALF OF THE FIBER -TO-THE-HOME COUNCIL

JANUARY 5, 2009

ECONOMIC EFFECTS OF TAX INCENTIVES FOR BROADBAND


INFRASTRUCTURE DEPLOYMENT

EXECUTIVE SUMMARY

Investments in next generation broadband infrastructure, such as fiber to the home networks,
generate both immediate and long-term benefits for the U.S. economy. In the short run, increased
capital investment leads directly to increased employment and output. In the longer run, the
rapid deployment of affordable broadband services transmitted over next generation
infrastructure is essential to U.S. competitiveness. Tax incentives to encourage deployment of
these high-speed broadband services therefore represent an efficient mechanism for increasing
both short-term economic growth and long-run economic competitiveness.

We analyzed four proposals: (1) 100 percent expensing of investments made in next generation
broadband networks - defined as those networks capable of delivering at least 100 megabits
downstream and 20 megabits upstream; (2) 50 percent expensing of broadband investments in
rural and underserved areas capable of delivering at least 5 megabits downstream and 1 megabit
upstream; (3) tax-credit bonds for private investments in next generation broadband
infrastructure; and (4) tax-credit bonds for public sector investments in next generation
broadband infrastructure.

Our results demonstrate that each of these proposals would generate substantial net benefits for
the U.S. economy. Specifically:

• The two tax credit bond proposals would have the largest impact on the economy, generating
more than $30 billion in new investment in next generation broadband infrastructure and
more than $ 100 billion in additional GDP over the next three years, and directly creating
approximately 215,000 net new jobs in each of the next three years.

• The expensing proposals - which are far less "expensive" in terms of forgone tax revenues -
would also have significant effects on investment, growth and employment, generating up to
$6 billion in new investment and over $18 billion in increased GDP over the next threes ears,
and directly creating approximately 37,000 net new jobs in each of the next three.

• All of the proposals represent efficient mechanisms for stimulating economic activity and
employment. Even ignoring the offsetting tax revenues that would result from increased
employment and economic activity, and counting only direct employment effects, the tax
expenditure per new job created is between $50,000 and $57,000 for the three proposals
involving next generation networks, and approximately $71,000 for the rural/underserved
proposal.

• The proposals would significantly increase next generation broadband availability overall
and current generation availability in rural and underserved areas, reduce broadband prices

"!5T ii nr -tc t5c


eau^aciF

(as measured by price per megabit), increase broadband penetration, and thus result in
substantial indirect effects on productivity, growth and employment. Under the two
expensing proposals, for example, up to 6.6 million additional homes would be passed by
fiber to the home type networks, and broadband service would become available to 4.0
million homes in rural and underserved areas that do not have broadband access.

• Increased broadband penetration resulting from lower prices and increased availability would
result in additional "indirect" job creation. For example, the two tax-credit bond proposals
would result in a sustained increase in employment of nearly 360,000 new jobs.

Finally, it should be noted that these proposals, if adopted, would affect economic activity almost
immediately. Private sector firms are already in the field deploying new broadband infrastructure
and have the ability to further accelerate planned deployments. The temporary nature of the four
proposals analyzed here would give these firms very strong incentives to "front-load" investment
activities that might otherwise be stretched out over the course of many years (especially in view
of the current downturn in economic activity).

The results of our analysis are summarized in Table 1, which is reproduced below.

TABLE 1: SUMMARY OF ANNUAL DIRECT ECONOMIC EFFECTS ON JOBS AND OUTPUT, 2009-2011
Private
50% Expensing for Sector Public
5/1 Mbps Tax Sector Tax
00% Expensing (Rural/Underserved Credit Credit
for 100/20 Mbps Areas only) Bonds Bonds
Direct Effects
- Output ($Billion, 2009-2011 total) 5.214 - 15.334 1.051 - 3.091 93.878 9.388
- Jobs (Annual Increase) 10,965 - 32,250 1,840 - 5,413 197,437 19,744
Forgone Tax Revenues over Investment
0.583 -1.715 0.131 - 0.386 11.178 0 . 985
Life ($Billion)
$ Forgone Tax Revenue per Direct
$53 , 182 $71 , 229 $56 , 616 $49 , 889
Effect Job
Direct Jobs per $Million Forgone Tax
18.804 14.039 17.663 20.045
Revenue

ii

CONTENTS

1. Introduction ....................................................................................................................... l

II. The Impact of Tax Incentives on Investment and the Economy .................................. 2

III. Proposals Analyzed ........................................................................................................... 2

A. Expensing Proposals ............................................................................................... 2

B. Private Sector Tax-Credit Bond Proposal ............................................................... 3

C. Public Sector Tax-Credit Bond Proposal ................................................................ 3

IV. Analysis of Direct Effects ................................................................................................. 4

A. Description of Data ................................................................................................. 4

B. Direct Effects of 100/20 Mbps Expensing (100 percent) ..................................... 11

C. Direct Effects of 511 Mbps Expensing (50 percent) ............................................. 11

D. Direct Effects of Private Sector Tax-Credit Bonds ............................................... 12

E. Direct Effects of Public Sector Tax-Credit Bonds ................................................ 12

V. Impact on Tax Revenues ................................................................................................13

VI. Analysis of Indirect Effects ............................................................................................15

A. Methodology and Assumptions ............................................................................ 16

B. Results of Indirect Effects Analysis ...................................................................... 18

VII. Conclusion .......................................................................................................................18


I. INTRODUCTION

1. We have been asked by the Fiber-to-the-Home Council (FTTH Council) to


analyze the economic impact of proposed tax incentives for broadband deployment. We analyze
four specific proposals:

A) Immediate expensing of 100 percent of investments providing 100 megabit


downstream/20 megabit upstream service to any area in the United States for
three years (2009-2011),

B) Immediate expensing of 50 percent of investments providing 5 megabit


downstream/1 megabit upstream service to viral and underserved areas in the
United States for three years (2009-2011),

C) Issuance by private sector entities of up to $10 billion in tax-credit bonds per year
over the next three years (2009-2011) to fiend investments on broadband
deployments providing 100 megabit downstream/20 megabit upstream service to
any area in the United States; and

D) Issuance by public sector entities of up to $1 billion in tax-credit bonds per year


over the next three years (2009-2011) to fund investments on broadband
deployments providing 100 megabit downstream/20 megabit upstream service to
any area in the United States.

2. Each of these proposals will generate significant benefits for the U.S. economy,
measured both in increased GDP and increased employment. GDP and employment will increase
over the next three years because of the increased investments by broadband providers resulting
from the tax relief ("direct effect"). Table 1 shows the economic impact of each of the four
proposals.

TABLE 1: SUMMARY OF DIRECT ECONOMIC EFFECTS ON JOBS AND OUTPUT, 2009-2011


Private
50% Expensing for Sector Public
5/1 Mbps Tax Sector Tax
00% Expensing (Rural/Underserved Credit Credit
for 100/20 Mbps Areas only) Bonds Bonds
Direct Effects
- Output ($Billion, 2009-2011 total) 5.214 - 15.334 1.051 - 3.091 93.878 9.388
- Jobs (Annual Increase) 10,965 - 32,250 1,840 - 5,413 197,437 19,744
Forgone Tax Revenues over Investment
0 . 583 - 1 . 715 0 . 131 - 0 . 386 11 . 178 0 . 985
Life ($Billion)
$ Forgone Tax Revenue per Direct
$53 , 182 $71 , 229 $56 , 616 $49 , 889
Effect Job
Direct Jobs per $Million Forgone Tax
18.804 14.039 17.663 20.045
Revenue

3. As Table 1 shows, the impact on economic output from 2009 to 2011 ranges from
$1.051 billion for the 50 percent expensing proposal to $93.878 billion for the private sector tax-
credit bonds. The increase in average annual employment ranges from 1,840 net new jobs for the
50 percent expensing proposal to 197,437 net new jobs for the private sector tax-credit bond
proposal.

4. Table 1 also shows the forgone tax revenues from each proposal. Our estimates of
forgone tax revenues represent only the direct effect of each policy, and do not account for
offsetting revenues resulting from increased incomes for suppliers of the inputs for broadband
deployment (e.g., income taxes resulting from increased employment). Our estimates of the
forgone tax revenues over the 15-year depreciable life of the investments made from 2009 to
2011 range from $131 million to $11.2 billion for each of the four proposals. Thus, from 2009 to
2011, each of the four proposals will sustain an average of between 14 and 20 net new jobs per
million dollars of forgone tax revenue as a result of the direct effect of increased broadband
capital expenditures.

II. THE IMPACT OF TAX INCENTIVES ON INVESTMENT AND THE ECONOMY

5. Investment tax incentives affect the economy by reducing the after-tax cost of
investment and thus increasing the effective rate of return on investment (ROI) from what it
would be in the absence of the tax incentive. As a result, firms choose to make investments that
would otherwise be uneconomic, and the overall amount of investment in the economy increases
accordingly.

6. By increasing investment, investment tax incentives have a direct effect on


employment and output. The direct effects are jobs and economic activity created as a direct
result of increased outlays for equipment, increased employment for installation, and associated
expenses (e.g., jobs resulting from increased purchases of equipment needed for installation,
such as bucket trucks and construction equipment). The most authoritative and generally
accepted means of estimating the direct effect of increased investment is the RIMS II model,
developed by the Bureau of Economic Analysis.

III. PROPOSALS ANALYZED

7. We analyzed four specific proposals. In this section, we briefly describe each.

A. Expensing Proposals

8. Expensing (or accelerated depreciation) affects the after-tax cost of investment by


allowing a firm to deduct from its taxable earnings the full amount spent on the investment,
rather than stretching that deduction out based on the depreciation schedule for that investment.
The after-tax cost of the investment is thus reduced by the difference between the value of the
tax deduction taken in year one, on the one hand, and the present value of the flow of tax
deductions that would otherwise be taken over the life of the equipment. The impact of

expensing thus depends on the depreciation life (for tax purposes) of the eligible investment, and
on the applicable tax rate.

9. 100/20 Mbps: The specific expensing proposal we were asked to analyze would
allow for immediate expensing of 100 percent of investments made over three years (2009-2011)
that provide 100 megabit downstream/20 megabit upstream service to any area in the United
States.

10. 5/1 Mbps Rural & Underserved: The second specific expensing proposal we
were asked to analyze would allow for immediate expensing of 50 percent of investments made
over three years (2009-2011) that provide 5 megabit downstream/1 megabit upstream service to
rural and underserved areas of the United States.

B. Private Sector Tax-Credit Bond Proposal

11. Tax-credit bonds are debt instruments that qualify bondholders to receive tax
credits from the U.S. Treasury, effectively reducing the bondholders' tax liability by an amount
equal to the tax credit. As a result, the yield required to sell such bonds at par is reduced by the
value of the tax credit to the bonds' purchasers.

12. The tax-credit bond proposals we were asked to analyze call for the Secretary of
the Treasury to establish tax credits which allow issuers to sell the bonds at a zero coupon rate.
Thus, bondholders would receive tax credits equal to the amount they would have received in
interest had the bonds been sold without the tax credit. Under this proposal, private sector
entities would be able to borrow up to $10 billion in tax-credit bonds per year over the next three
years (2009-2011) to fund investments on broadband deployments providing 100 megabit
downstream/20 megabit upstream service to any area in the United States.

C. Public Sector Tax-Credit Bond Proposal

13. Our analysis of the public sector tax-credit bond proposal is similar to the analysis
of the private sector tax-credit bond proposal. The public sector proposal would allow for public
sector entities to partner with private sector entities in the deployment of broadband. This access
to lower-cost funding would induce a firm to invest more in broadband deployments than it
would absent the tax incentive.

14. The specific tax-credit bond proposal we were asked to analyze would allow for
the issuance by public sector entities of up to $1 billion in tax-credit bonds per year over the next
three years (2009-2011) to fund investments on broadband deployments providing 100 megabit
downstream/20 megabit upstream service to any area in the United States. The FTTH Council
believes that providing $1 billion per year in bonds is appropriate for this program, as compared
to the larger $10 billion in bonds for the private sector program described above, because to date
public sector entities have been involved more selectively in deploying broadband infrastructure
and because, at least for municipalities, they are limited in the scale of their deployments by the

geographic limitations of their jurisdictions. In addition, public sector entities typically take
substantially longer to deploy broadband networks than private sector entities - in some recent
cases, approximately three years from proposal to groundbreaking.

IV. ANALYSIS OF DIRECT EFFECTS

15. Our analysis of each proposal entailed estimating the direct effects of increased
spending resulting from the tax incentives. All of our models include various baseline
assumptions related to the number of homes passed and served by broadband technology and the
investment required to deploy and maintain broadband lines. For estimates relating to modeling
broadband service of 100/20 Mbps, we use historical data and projections on fiber-to-the-home
(FTTH), the most prevalent form of techmology currently used to achieve the speeds required to
meet the tax incentive thresholds. For estimates relating to modeling broadband service of 5/1
Mbps, we use historical data and projections on cable and digital subscriber line (DSL)
broadband service.

A. Description of Data

• 100/20 Mbps Service (FTTH)

16. We use historical data and an average of forecasts of homes passed and homes
served by FTTH through 2013 from RVA Market Research (RVA). Dividing RVA's forecasts of
the number of homes passed and homes served by Morgan Stanley's forecasted number of
households through 2011 1 yields annual fiber penetration rates and adoption rates.

17. To estimate the cost to deploy and serve a home with fiber, we use estimates from
a proposal for fiber deployments for the city of Portland, Oregon.Z According to the 2007
proposal by Uptown Services, capital expenditures per home passed with FTTH were $765 in
outside plant build-out costs. Uptown Services estimated that subscriber capital investments,
which would include optical network terminals (ONTs), drop cables, connectors, ONT power
supply, and set top boxes would be between $667 (without digital video recorder (DVR)) and
$817 (with DVR) per new subscriber. Therefore, we assume that the investment required to pass
a home with fiber is $765 and the additional investment required to serve a home is $742
(average of $667 and $817) in 2007. After 2007, we assume a 5 percent annual decrease in the
investments required to pass and serve a home with fiber.

18. Finally, we assume that 100 percent of forecasted fiber capital expenditures would
meet the speed limits necessary for eligibility for the tax expensing and tax-credit bond proposals

i Simon Flannery, Benjamin Swinburne, David Gober, Daniel Gaviria, & Chad Harris, Morgan Stanley,
Cable/Sat & Telecom Broadband Outlook: Online Usage Growth Favors Cable, DirecTV Remains HD Leader (July
18, 2008), at Ex. 26.
2 Uptown Services, LLC, "Phase 2 Business Case for a Community Fiber Network, Prepared for the City of
Portland by Uptown Services, LLC," Nov. 2007, at 25.

because RVA's forecasts are based on deployments that can meet the 100/20 Mbps speed
thresholds.

• 5/1 Mbps Service (Cable and DSL)

19. To estimate the number of homes passed by 5/1 Mbps service without the tax
incentive, we use data from Morgan Stanley's forecast of residential cable and DSL subscribers
through 2011.3 Morgan Stanley presents forecasts of broadband subscribers by cable and
combined DSL+Fiber service. We estimate DSL subscribers by subtracting the RVA forecasts of
fiber subscribers from Morgan Stanley's forecast of DSL+Fiber subscribers. Morgan Stanley
forecasts the number of homes passed for cable broadband services, but does not forecast the
number of homes passed by DSL service.4 We assume that if a home is not passed by cable
broadband, then it is not passed by DSL. Based on Morgan Stanley's forecasts of homes passed
by cable broadband and total households, an average of 7.3 million homes (equal to 6.0 percent
of all households) will not be passed by broadband from 2009 through 2011 without the
proposed tax incentives.

20. To estimate the cost to deploy and serve a rural or underserved home with cable,
we use capital expenditure estimates of cable deployment from Morgan Stanley. Both cable and
DSL broadband require three types of capital expenditures: (1) deployment capital expenditures,
or investment in upgrading networks; (2) expenditures on customer premises equipment (CPE),
such as modems; and, (3) maintenance capital expenditures. Morgan Stanley forecasts capital
expenditures through 2012 on rebuilds and upgrades of cable networks for advanced services,
including broadband, digital cable, and telephony, per basic subscriber, expenditures on CPE per
net additional broadband subscriber, and maintenance capital expenditures on broadband per
existing subscriber.5 Morgan Stanley's average forecasted estimates for cable broadband capital
expenditures between 2009 and 2011 were (1) $100 per new subscriber in CPE, (2) $3 per total
basic cable subscriber in investments to upgrade service to broadband capability, and (3) $20 in
maintenance investments per cable broadband subscriber. Morgan Stanley's $3 estimate of
capital expenditures for upgrades is low because it is an average expenditure for all basic
subscribers, not just those who are being upgraded. Using the estimates of basic subscribers and
new homes passed, Morgan Stanley's estimates show that the cost to upgrade service is $213 per
new home passed in 2009. This is the estimate we use for estimating the cost to upgrade cable
for providing broadband service to a rural or underserved customer in 2009, and we assume that
this cost declines by 5 percent each year.

3 Simon Flannery, Benjamin Swinburne, David Gober, Daniel Gaviria, & Chad Harris, Morgan Stanley,
Cable/Sat & Telecom Broadband Outlook: Online Usage Growth Favors Cable, DirecTV Remains HD Leader (July
18, 2008), at Ex. 26.
4 Id. at Ex. 23.
5 Richard Bilotti, Benjamin Swinburne, & Megan Lynch, Morgan Stanley, Truth, Lies and Truck Rolls:
Understanding Product Profitability 8 (Oct. 4, 2002).

21. To estimate the cost to deploy and serve a rural or underserved home with DSL,
we use a Bear Stearns report that forecasts DSL deployment and CPE capital expenditures per
new customer and DSL maintenance capital expenditures per existing customer through 2005.6
After 2005, we assume that each line-item expenditure in Bear Stearns's capital expenditure
forecasts decreases by 10 percent annually.

22. Additionally, Bear Stearns presents capital expenditure forecasts for two
categories of customers: those located within 18,000 feet of a service provider's central office
and those beyond 18,000 feet. The cost to deploy DSL to a customer beyond 18,000 feet is
higher than the cost to deploy DSL to a customer within 18,000 feet. Because the deployments
under the proposed expensing legislation would be made to rural and underserved areas, we use
the higher deployment costs from the Bear Stearns report for customers beyond 18,000 feet of
the service provider's central office when estimating increased capital expenditures in rural and
underserved areas.

23. We estimated capital expenditures on DSL and cable broadband without the tax
incentive by multiplying the various components of capital expenditures by the relevant factor-
new homes passed, gross subscriber additions, or existing subscribers. In calculating gross
subscriber additions, we assume that 75 percent of new fiber subscribers each year are former
DSL subscribers and that 15 percent of new fiber subscribers each year are former cable
broadband subscribers. Thus, the number of gross subscriber additions is equal to Morgan
Stanley's forecasted net subscriber additions plus the number of DSL and cable broadband
subscribers we assume switched to fiber.

24. Finally, our analysis required an assumption about the share of cable and DSL
capital expenditures that would be eligible for the proposed tax incentive. We assume that all
expenditures would meet the speed threshold. We assume that 63 percent of the forecasted cable
and DSL capital expenditures would meet the rural and underserved area qualifications for 50
percent tax expensing. In earlier work by two of the authors of this study, we estimated the rural-
underserved share of homes passed to apply to the total capital expenditures. We assumed that
this share was equal to the percentage of households served by at least four broadband providers
that reside in zip codes that are rural or underserved. To calculate this figure, we first classified
zip codes as being rural or underserved if at least 50 percent of the Census tracts intersecting the

6 Robert Fagin, Bear Stearns, Wireline Services: The DSL Report: Demystifying the Economics of Digital
Subscriber Line, Exhibit 6 (Sept. 2002). Our estimates differ from Bear Steams in that we calculate maintenance
capital expenditures per existing DSL subscriber in a year to be equal to 15 percent of the sum of deployment
equipment, incremental bandwidth, and ATM switching capacity capital expenditures per newly deployed DSL
customer in that year. This estimate of maintenance capital expenditures produces results that more closely match
the DSL maintenance capital expenditures per line estimated by other analysts. See, e.g., Douglas S. Shapiro, Banc
of America Securities, Broadband Brief. DSL Economics, Game Theo7y and What Happens to Broadband Pricing
Next 4 (Sept. 8, 2003). Banc of America estimates annual DSL maintenance capital expenditures per subscriber to
be $46 in 2003 and $36 in 2008. Using our methodology and Bear Steams's estimates of deployment capital
expenditures, we estimate annual DSL maintenance capital expenditures per existing subscriber to be $42 in 2009.

boundaries of the zip code are classified as rural or underserved according to the definitions used
in previous legislation that is consistent with the current proposal. We then matched these zip
codes with June 2000 and June 2002 data from the FCC that shows the number of firms
providing broadband service by zip code. We estimated that 18 percent of the households in zip
codes that were served by at least four high-speed service providers were in rural or underserved
zip codes in June 2000. We calculated that this share grew to 28 percent in June 2002a 10-
percentage-point increase over two years. Based on this increase, we assume that the share
continued to increase by 5 percentage points each year. Therefore, we assume that, absent the tax
incentive, 63 percent of capital expenditures on current generation technology in 2009 will be
spent on deployments in rural and underserved areas, 68 percent will go to rural and underserved
areas in 2010, and so forth. When capital expenditures increase as a result of the tax incentive,
we assume that 100 percent of the additional DSL and cable customers and homes passed
resulting from the tax incentive would be located in rural and underserved areas.

• Multipliers

25. The incremental residential broadband capital expenditures that result from these
policies will have a multiplicative effect on the economy when the economy is at less than full
employment, as it is today.' To estimate this multiplicative effect, we use the most recent RIMS
II multipliers on detailed industries by NAICS code, based on 1997 national benchmark input-
output data and 2006 regional data. Broadband deployment requires capital spending on
equipment and construction. Therefore, we use multipliers for telephone apparatus
manufacturing, fiber optic cable manufacturing, and construction. Table 2 shows the industry
multipliers we use and the weights assigned to each industry to estimate the average multiplier
for broadband investment.8

7 The multiplier is a standard principle in the macroeconomics literature. See, e.g., RUDIGER DORNBUSCH &
STANLEY FISCHER, MACROECONOMICS 66 (McGraw Hill 6th ed. 1994). Richard Kahn first introduced the multiplier
concept as an "employment multiplier." See Richard F. Kahn, The Relation of Home Investment To Employment, 41
ECON. J. 173, 173-98 (1931). John Maynard Keynes expanded upon this concept by introducing the "investment
multiplier," which is the multiplier used in our analysis. See John Maynard Keynes, A GENERAL THEORY OF
EMPLOYMENT, INTEREST, AND MONEY 115 (Harcourt Brace & Co. 1964) (1936).
8 U.S. Department of Commerce, Bureau of Economic Analysis, Regional Input-Output Modeling System
(RIMS II), Table 1.5 (2008). Multipliers are based on the 1997 Benchmark Input-Output Table for the Nation and
2006 regional data. These industries approximately match the expenditures made to deploy and connect broadband
more closely than any other multiplier category. According to the 1997 NAICS definition, industry 334210
(Telephone apparatus manufacturing) consists of "[e]stablishments primarily engaged in manufacturing wire
telephone and data communications equipment. These products may be standalone or board-level components of a
larger system. Examples of products made by these establishments are central office switching equipment, cordless
telephones (except cellular), PBX equipment, telephones, telephone answering machines, and data communications
equipment, such as bridges, routers, and gateways." Industry 335921 (Fiber optic cable manufacturing) consists of
"[e]stab lishments primarily engaged in manufacturing insulated fiber-optic cable from purchased fiber-optic strand."
Industry 230000 (Construction) includes, among other types of construction establishments, "[e]stablishments
primarily responsible for the entire construction (i.e., new work, reconstruction, or repairs) of electric power and
communication transmission lines and towers, radio and television transmitting/receiving towers, cable laying, and

TABLE 2: MULTIPLIERS FOR BROADBAND CAPITAL EXPENDITURES


Final
Final Demand:
Demand: Employment
Output (Jobs per TTH able SL ireless
(GDP $ per Million $ Industry Industry Industry Industry
AICS Industry Invested) Invested) Weight Weight Weight Weight
334210 Telephone apparatus
2.6424 11.7592 30% 80% 80%
manufacturing
334220 Broadcast and wireless
2.8309 13.7828 0% 0% 0% 93%
communications equipment
335921 Fiber optic cable manufacturing 3.0284 14.4066 20% 0% 0%

230000 Construction 3.4617 26.6692 50% 20% 20% 7%


FTTH Weighted Average Multiplier 3.1293 19.7437
Cable Weighted Average Multiplier 2.8063 14.7412
DSL Weighted Average Multiplier 2.8063 14.7412
Wireless Weighted Average Multiplier 2.8739 14.6618

26. According to Uptown Services, a majority (54 percent) of capital spending


required in outside plant build-out for FTTH is spent on construction.9 This heavy reliance on
construction for FTTH is due in large part to the burying of new infrastructure in the ground.
Construction is given a larger weight for FTTH than for DSL, cable or wireless because much of
the infrastructure over which cable (e.g., conduits and HFC cable), DSL (i.e., copper), and
wireless (i.e., towers) already exists and does not require new construction. As Table 2 shows,
the multipliers for the construction industry are substantially larger than the multipliers for the
other three industries. For example, $1 million of investment in FTTH deployment will result in
almost 20 jobs, whereas a dollar of investment in wireless broadband will result in fewer than 15
jobs.10 This is largely due to our estimate that only 7 percent of wireless broadband capital
expenditures go to the construction industry. I I

cable television lines; (2) establishments identified as power and communication transmission line construction
management firms; and (3) establishments identified as special trade contractors engaged in activities primarily
related to power and communication transmission line construction." Industry 334220 (Broadcast and wireless
communications equipment) includes "establishments primarily engaged in manufacturing radio and television
broadcast and wireless communications equipment. Examples of products made by these establishments are:
transmitting and receiving antennas, cable television equipment, GPS equipment, pagers, cellular phones, mobile
communications equipment, and radio and television studio and broadcasting equipment." See U.S. Census Bureau,
1997 NAICS and 1987 SIC Correspondence Tables, available at http://www.census.gov/epcd/www/naicstab.htm.
9 Uptown Services, LLC, "Phase 2 Business Case for a Community Fiber Network, Prepared for the City of
Portland by Uptown Services, LLC," Nov. 2007, at 25.
10 The employment multipliers in Table 2 represent the effect of investments on jobs within the United States.
As Table 2 shows, the employment multiplier for the construction industry is approximately twice as large as the
multipliers for the other industries. This difference is due in large part to the concentration of construction jobs

27. The multiplier specific to the industries shown in Table 2 translates the effect of
broadband capital spending on U.S. employment and gross domestic product (GDP). The
multiplicative effect occurs because higher expenditures on broadband deployment-equivalent
to higher demand for construction and the products of equipment manufacturers-causes the
equipment manufacturers and construction firms to hire more employees to meet the increased
demand. The equipment manufacturers' incomes and construction firms' incomes increase as
well due to the increased expenditures, which, according to the consumption function, will
increase their consumption as well. The increased consumption of equipment manufacturers and
construction firms will in turn increase the income and employment of their suppliers. The
income and employment of those suppliers will then increase, and so on.

28. Table 2 shows that a one-million dollar increase in the final demand for
communications infrastructure investment by fiber broadband providers would create nearly 20
new jobs nationally. The timeframe over which employment would increase is debatable. In
most cases, the BEA considers one year to be the appropriate time horizon for its multipliers to
have achieved full effect. 12 Other economists have estimated that at least two years may be
required for incremental investment to achieve its full impact on the economy. 13 The multiplier
effect is most fully realized when there is substantial excess capacity, during economic
recessions or sharp declines in specific sectors. Because the economy is in the midst of a
recession, 14 excess capacity exists. Accordingly, our estimates of the multiplier effect of
increased capital expenditures reasonably capture the effect that increased capital spending by
broadband providers would have on the U.S. economy.

within the United States relative to the other industries. For example, a dollar spent on telephone equipment may be
spent in a factory overseas, resulting in an increase in foreign employment. Construction, on the other hand, is a
local industry that requires U.S.-based workers to perform its essential functions. Therefore, a dollar spent on the
construction industry will lead to more U.S. job growth than a dollar spent on other industries in which much of the
main output is produced overseas.
According to a report by the WiMax Forum, 7 percent of the 5-year capital expenditures on WiMax
deployment in rural areas (the only areas in which WiMax would be eligible for any of the tax proposals we
analyze) would be spent on "site acquisition and civil works." This component appears to be focused more on the
construction industry, whereas the other components of capital expenditures in the WiMax report are focused on
equipment such as CPE, base station equipment, and base station backhaul. WiMax Forum, Business Case Models
for Fixed Broadband Wireless Access based on WiMAX Technology and the 802.16 Standard (Oct. 10, 2004), at 20.
12 U.S. DEPARTMENT OF COMMERCE, BUREAU OF ECONOMIC ANALYSIS, REGIONAL INPUT-OUTPUT MODELING
SYSTEM REGIONAL MULTIPLIERS: A USER HANDBOOK FOR THE REGIONAL INPUT-OUTPUT MODELING SYSTEM
(RIMS 11), at 8 (Mar. 1997).
1' See, e.g., OLIVER BLANCHARD, MACROECONOMICS 72-73 (Prentice Hall 1997).
14 In its December 2008 announcement that the current recession began in December 2007, the National
Bureau of Economic Research noted that payroll employment had declined in every month since December 2007.
See National Bureau of Economic Research, Determination of the December 2007 Peak in Economic Activity,
available at http://www.nber.org/cycles/dec2008.html (Dec. 11, 2008).

• Tax-Credit Bonds

29. As discussed above, the tax-credit bond proposals we examine call for the
Secretary of the Treasury to establish tax credits that allow the bonds to be sold at a zero coupon
rate, i.e., providing the eligible borrowers with interest-free financing for the eligible projects.
We assume that these terms are sufficiently attractive that the bonds would be utilized up to the
specified limits, i.e., $10 billion annually for private tax-credit bonds and $1 billion annually for
public tax-credit bonds. We assume that an equal amount of $10 billion in private sector tax-
credit bonds are issued annually. Because public sector projects may take longer to begin than
private sector projects, we assume that no public sector tax-credit bonds are issued in 2009, $1
billion is issued in 2010, and $2 billion is issued in 2011 (pursuant to the catch-up provisions of
the tax proposal in which the $1 billion limit in public sector tax-credit bonds for a subsequent
year is increased if the full amount of the bonds are not issued in a given year).

30. Further, because the proposal calls for the bonds to be used only to finance
projects approved by state public utility commissions (for private bonds) and state governments
and the U.S. Department of Commerce (for public bonds), we assume that 100 percent of the
investment that results is incremental, i.e., used for projects that would not otherwise have been
undertaken. Hence, we assume that the effect of both tax-credit bond proposals is to increase
total investment in FTTH projects by an average of $11 billion annually for three years.

• Other Assumptions

31. Each expensing proposal lowers the after-tax cost of the goods and services
purchased through broadband provider's capital investments. Under the 100 percent tax
expensing proposal, expenditures are expensed completely in the year they are made. Without
the expensing proposal, those expenditures would have been expensed over several years
according to the appropriate depreciation schedule. To estimate the effective decrease in cost
resulting from the tax expensing proposal, we estimate the net present value (NPV) of the
forgone tax savings in future years for the broadband provider resulting from the immediate
expensing of capital in year one under the proposal. We assume that the investment is 15-year
depreciable property, and the taxpayer follows a half-year convention and applies a 150 percent
declining balance depreciation method. Therefore, from a $100 investment, we deduct $5.00 for
normal first year depreciation. This leaves $95 to be deducted under broadband expensing. We
then determine the NPV of a $95 tax deduction, which we estimate at $33.25, assuming a 35
percent corporate tax rate. Next, we reduce $33.25 by the NPV of the year 2-15 depreciation
deductions that would have been available in the absence of broadband expensing, equal to
$18.17 (using a weighted average cost of capital (WACC) of 10 percent). Reducing $33.25 by
$18.17, the remaining $ 15.08 would be the benefit of 100 percent broadband expensing, equating
to 15.08 percent of the total investment. Using a similar calculation for the rural and underserved
area tax incentive, the benefit of 50 percent broadband expensing would be 7.54 percent of the
total investment.

10

32. To estimate changes in capital expenditures resulting from the lower after-tax cost
of the products and services purchased by broadband providers due to the expensing proposals,
we assume that the elasticity of the broadband providers' demand for those products and services
is between -0.85 and -2.5. With an elasticity of demand of -0.85, a reduction in the broadband
provider's cost of expenditures of 1 percent will increase its demand for those products and
services by 0.85 percent. Likewise, an elasticity of -2.5 indicates that a reduction in the
broadband provider's cost of expenditures of 1 percent will increase its demand for those
products and services by 2.5 percent.

33. Finally, all of our estimates assume continuation of the current regulatory
environment for broadband deployment and access. Any additional regulations, such as open
access rules for FTTH, would decrease our estimates of broadband investments and their direct
effects on economic output and employment.

B. Direct Effects of 100/20 Mbps Expensing (100 percent)

34. Table 3 shows our estimates of the direct effect of increased capital expenditures
in FTTH if the 100/20 Mbps broadband expensing proposal is implemented for 2009-2011.

TABLE 3: DIRECT ECONOMIC EFFECT OF 100/20 MBPS TAX EXPENSING PROPOSAL


2009 2010 2011 Total
FTTH Capital. Expenditures before
4.031 4.314 4.651 12.996
Tax Proposal ($Billion)
FTTH Capital Expenditures after
4.548 - 5.551 4.867 - 5.940 5.247 - 6 .405 14.662 - 17.896
Tax Proposal ($Billion)
Increase in Capital Expenditures
0.517 - 1.520 0.553 - 1.627 0.596 - 1.754 1.666 - 4.900
(Willion)
Direct Effect on GDP ($ Billion) 1.617 - 4.757 1.731 - 5.090 1. 866 - 5.488 5.214 - 15.334
Direct Effect on Employment
10,204 - 30,012 10,919 - 32,114 11,772 - 34 ,624 10,965 - 32,250
(Jobs)

35. As Table 3 shows, we estimate that the 100/20 Mbps expensing proposal will
increase capital expenditures on FTTH by between $1.7 billion and $4.9 billion from 2009 to
2011. This increase will directly result in an increase in GDP of between $5.2 billion and $15.3
billion over the three years. On average over the three years, the increased investment will
maintain an additional 10,965 to 32,250 jobs per year.

C. Direct Effects of 5/1 Mbps Expensing (50 percent)

36. Table 4 shows our estimates of the direct effect of increased capital expenditures
in FTTH if the 511 Mbps broadband expensing proposal is implemented for 2009-2011.

11

TABLE 4: DIRECT ECONOMIC EFFECT OF 5/1 MBPS TAX EXPENSING PROPOSAL


2009 2010 2011 Total
Cable/DSL Capital Expenditures before
3.076 2 .931 2.619 8.626
Tax Proposal ($Billion)
Cable/DSL Capital Expenditures after
3.200 - 3.441 3.059 - 3.307 2.742 - 2.980 9.001 - 9.728
Tax Proposal ($Billion)
Increase in Capital Expenditures
0.124 - 0.365 0.128 - 0.376 0.123 - 0.360 0.375 - 1.102
($Billion)
Direct Effect on GDP ($Billion) 0.349 - 1.025 0.359 - 1.054 0.344 - 1.012 1.051 - 3.091
Direct Effect on Employment (Jobs) 1,831 - 5,385 1,883 - 5,539 1,807 - 5,314 1,840 - 5,413

37. As Table 4 shows, we estimate that the 5/1 Mbps expensing proposal will increase
capital expenditures on cable and DSL by between $375 million and $1.1 billion from 2009 to
2011. This increase will directly result in an increase in GDP of between $1.1 billion and $3.1
billion over the three years. On average over the three years, the increased investment will
maintain an additional 1,840 to 5,413 jobs per year.

D. Direct Effects of Private Sector Tax-Credit Bonds

38. Table 5 shows the direct effect on the economy of $10 billion in additional
investment on FTTH each year from 2009 to 2011 that results from the proposed private sector
tax-credit bonds.

TABLE 5: DIRECT ECONOMIC EFFECT OF PRIVATE SECTOR TAX-CREDIT BONDS


2009 2010 2011 Total
FTTH Capital Expenditures before Tax
4.031 4.314 4 . 651 12 . 996
Proposal ($Billion)
FTTH Capital Expenditures after Tax
14.031 14.314 14.651 42 . 996
Proposal ($Billion)
Increase in Capital Expenditures
10.000 10.000 10.000 30.000
($Billion)
Direct Effect on GDP ($ Billion) 31.293 31.293 31.293 93.878
Direct Effect on Employment (Jobs) 197,437 197,437 197,437 197,437

39. As Table 5 shows, we estimate that the private sector tax-credit bond proposal
will increase capital expenditures on FTTH by $30 billion from 2009 to 2011. This increase will
directly result in a $93.9 billion increase in GDP over the three years. On average over the three
years, the increased investment will maintain an additional 197,437 jobs per year.

E. Direct Effects of Public Sector Tax-Credit Bonds

40. Table 6 shows the direct effect on the economy of the additional investment on
FTTH each year from 2009 to 2011 that results from the public sector tax-credit bonds.

12

TABLE 6: DIRECT ECONOMIC EFFECT OF PUBLIC SECTOR TAX-CREDIT BONDS


2009 2010 2011 Total
FTTH Capital Expenditures before Tax
4.031 4.314 4.651 12.996
Proposal ($Billion)
FTTH Capital Expenditures after Tax
4.031 5.314 6.651 15 . 996
Proposal ($Billion)
Increase in Capital Expenditures
0.000 1.000 2.000 3.000
($Billion)
Direct Effect on GDP ($Billion) 0.000 3.129 6.259 9.388
Direct Effect on Employment (Jobs) 0 19,744 39,487 19,744

As Table 6 shows, 41. we estimate that the public sector tax-credit bond proposal will
increase capital expenditures on FTTH by $3 billion from 2009 to 2011. This increase will
directly result in a $9.4 billion increase in GDP over the three years. On average over the three
years, the increased investment will maintain an additional 19,744 jobs per year.

V. IMPACT ON TAX REVENUES

The impact on tax 42. revenues of the expensing proposals is dependent upon the
change in investment and the change in the timing of expensing. When a firm incurs additional
costs, it will be able to deduct those costs from its taxable income, thereby reducing the firm's
tax liability. Although changes in the timing of expensing will reduce tax revenues in the short-
run, (undiscounted) tax revenues over the life of the investment will be unchanged as long as the
amount invested does not change, and assuming the firm's marginal tax rate remains constant
over time.

43. We estimate the forgone tax revenues resulting from the proposed tax expensing
incentives by calculating the annual tax savings each firm enjoys both with and without the
incentive. A firm's tax savings in year t (tax,) from any investment originally made in year k
(invk) can be written as:

tax, = (invk * taxrate, * exp_rate,-k) + (invk * taxrate, * (1 -exp_ratek) * dep_rate, )

Part (a) represents the tax savings from an expensing rate of exp_rate in the year of the
investment. Part (b) represents the tax savings from the depreciation schedule where dep_rate, is
the percent of the investment remaining after expensing that is depreciated in year t. With no tax
incentives, the expensing exp_ratek rates is zero. With the 100 percent expensing proposal, invk
increases (relative to no tax incentive) and exp_ratek is 100 percent. With the 50 percent
expensing proposal, invk increases (relative to no tax incentive) and exp_ratek is 50 percent. We
assume a 35 percent marginal tax rate taxrater when estimating the tax revenue impact.

13

FTTH her to th47

44. The forgone tax revenues resulting from the tax-credit bond proposals are
functions of interest rates and tax rates. The effective interest rate on private borrowings under
the tax-credit bond proposal will reflect two factors. First, since interest on the bonds will
effectively be paid by the U.S. Treasury (in the form of tax credits), the default risk on the
interest component is effectively zero. Second, the default risk on the principal will be a function
of the risk characteristics of the issuers, which may range from major U.S. corporations to
smaller (and hence riskier) companies. For purposes of arriving at an estimate of the forgone tax
revenues, we assume that these two factors result in an effective interest rate of 4.14 percent,
equal to the average of the current yield for 10-year (5.5 percent) and 20-year (5.98 percent) A-
rated corporate bonds and the current yield for 10-year (2.16 percent) and 20-year (2.92 percent)
Treasury bonds. 15 The forgone tax revenues in each year until maturity resulting from the private
tax-credit bond proposal is equal to the effective interest rate multiplied by the amount issued.

45. The effective interest rate on public borrowings under the tax-credit bond
proposal will reflect two factors. First, since interest on the bonds will effectively be paid by the
U.S. Treasury (in the form of tax credits), the default risk on the interest component is effectively
zero. Second, the default risk on the principal is a function of the risk characteristics of the
issuers, which may range from state governments to local municipalities. For purposes of
arriving at an estimate of the forgone tax revenues, we assume that these two factors result in an
effective interest rate of 3.94 percent, equal to the average of the current yields on 10-year (4.81
percent) and 20-year (5.87 percent) A-rated municipal bonds and the current 10-year (2.16
percent) and 20-year (2.92 percent) Treasury yields as of December 22, 2008.16 The forgone tax
revenues in each year until maturity resulting from the public tax-credit bond proposal is equal to
the effective interest rate multiplied by the amount issued.

46. Table 7 shows the estimates of forgone tax revenues resulting from the tax
incentives and changes in capital expenditures. Table 7 shows both the impact on tax revenues
from 2009-2011 and the impact on revenues over the entire life of the investments made in 2009-
2011.17 Following the Joint Committee on Taxation, we do not discount the tax revenue cost of
the proposals. 18

15 Yahoo! Finance, Composite Bond Rates (http://finance.yahoo.com/bonds/composite bond rates); Federal


Reserve Board, Federal Reserve Statistical Release H.15, Selected Interest Rates
(.httf):,%'w.w_w_E_icier-ali-es4 }_e,;,co /re.leasesihl5 data. h.tin). Rates as of December 22, 2008.
Id.
" For tax-credit bonds, we calculate forgone tax revenues based on the Joint Committee on Taxation's usual
method of estimating tax effects over a ten -year budget window, rather than the entire life of the bonds.
18 Joint Committee on Taxation, Overview ofRevenue Estimating Procedures and Methodologies Used by the
Staff of the Joint Committee on Taxation (Feb. 2, 2005), at 12 (http ://www.house.gov/jct/x-1-05.pdf).

14

t, ,r f,-, .: ^C^?rp

TABLE 7: IMPACT OF TAX INCENTIVE PROPOSALS ON TAX REVENUES ($BILLIONS)


Tax Revenue
Reduction over Entire
2009-2011 Tax 15-Year Life of 2009-
Proposal Revenue Reduction 2011 Investments*
100% Expensing for 100/20 Mbps 4.506 - 5.638 0.583 - 1.715
50% Expensing for 5/1 Mbps (Rural/ Underserved Areas only) 1.363 - 1.508 0.131 - 0.386
Private Sector Tax-Credit Bonds 2.484 11.178
Public Sector Tax-Credit Bonds 0.158 0.985
* Forgone revenues in this column for tax-credit bonds represent interest payments over ten years from 2009-2018.

47. As Table 7 shows, the proposed 100 percent expensing proposal's effect on
increased capital expenditures reduces 2009 -2011 tax revenues by between $4.5 billion and $5.6
billion. The effect over the entire life of the increased investments made in 2009-2011 is between
$583 million and $1.7 billion for the 100 percent expensing proposal. The effect over the entire
life of the investments is smaller than the effect over 2009 to 2011 because the Treasury receives
more in tax revenues in the years after 2011 under 100 percent expensing than it does without
100 percent expensing. When 100 percent of an investment is expensed in the first year, there
will be no more investment to deduct from future years earnings. Without 100 percent expensing,
there are depreciated costs to deduct from earnings in every year through year 16 of the
investment.

48. By focusing only on firms' increased expenses, Table 7 overstates the true net
impact of the various tax proposals on tax revenues. We do not attempt to estimate the increase
in tax revenues that would result from the tax incentives in our analysis. For example, increased
employment through the direct effects would result in increased personal incomes, which would
result in increased income tax revenues. In addition, firms making the investments would see
their profits increase through greater consumption of their broadband services, which would
increase their corporate income taxes. The Joint Committee on Taxation (JCT) recently
estimated the combined cost of the 100/20 Mbps expensing provision and the 511 Mbps
expensing provision to be $72 million over ten years for a three-year provision. In making its
revenue impact calculations, the JCT generally accounts for income effects and other indirect
effects (discussed below) not included in Table 7 that increase tax revenues.

VI. ANALYSIS OF INDIRECT EFFECTS

49. This study focuses on the direct effects on the economy of each tax proposal. In
addition to these direct effects, the additional availability of broadband services will result in
increased adoption, which in turn will lead to increased productivity and demand for other goods
and services ("indirect effect"). The indirect effects of increased broadband investment result
from the productivity increases, price reductions, and related savings associated with increased
broadband adoption. The tax incentives at issue here would increase broadband adoption due to
both (a) increased broadband availability in rural and underserved areas and (b) reduced prices
and improved quality associated with the availability of more technologically advanced
broadband infrastructures generally. Our estimate utilizes reasonable assumptions regarding the

15

impact of increased availability, and applies the results of authoritative empirical research on the
impact of broadband adoption on employment to estimate these indirect effects. In this section,
we estimate the indirect effects from increased broadband adoption resulting from the increased
deployment that broadband providers will make as a result of the tax proposals discussed above.

A. Methodology and Assumptions

50. In our analysis of the direct effects of the tax proposals, we estimated the effects
of each proposal that result directly from increased investment in broadband infrastructure. The
ultimate effect of this investment, however, will be to increase the availability of next-generation
broadband services to households which already have some form of broadband available, and to
make broadband available in rural and underserved areas where broadband service is unavailable
today.

51. We model the adoption effect of increased high-speed broadband (100/20Mpbs)


as an effective reduction in the price, where price is measured as the monthly cost per
downstream megabit. 19 As shown in Table 8 below, the price per megabit for high speed services
is far lower than for slower DSL and cable connections.

TABLE 8: COMPARISON OF BROADBAND SPEEDS AND PRICES


Provider Service Typ e Download S peed Monthly Price $/Mb s
Cox Cable 768 Kbps $19.89 $25.90
Verizon DSL 768 Kbps $19.99 $26.03
Qwest DSL 1.5 Mb s $14.99 $9.99
AT&T DSL 1.5 Mbps $25.00 $16.67
Cox Cable 1.5 Mbps $29.99 $19.99
AT&T DSL 3.0 Mb s $29.95 $9.98
Verizon DSL 3.0 Mbps $29.99 $10.00
AT&T DSL 6.0 Mb s $35.00 $5.83
Comcast Cable 6.0 Mbps $57.95 $9.66
Qwest DSL 7.0 Mbps $24.99 $3.57
Comcast Cable 8.0 Mb s $67.95 $8.49
Cox Cable 9.0 Mbps $43.99 $4.89
Verizon FiOS 10 Mbps $47.99 $4.80
EarthLink Cable 10 Mbps $72.95 $7.30
Qwest DSL 12 Mbps $46.99 $3.92
Cox Cable 15 Mb s $56.95 $3.80
Verizon FI S 20 Mbps $57.99 $2.90
Verizon FiOS 50 Mb s $144.95 $2.90
Source: Company websites.

19 Price per megabit is a widely utilized measure of broadband pricing, since it captures the "quality" element
associated with higher speed services. See, e.g., http:!:'ww°w.oecd.org/sti:`ict:broadband.

16

52. We estimate conservatively that the effect of 100/20 Mbps fiber deployment in an
area already served by broadband is to reduce the average price of broadband in that area by
$3.67 per month per megabit, i.e., approximately the difference between an average of the
current pricing plans for 3 Mbps - 15 Mbps ($6.57/Mbps/month) and Verizon's current 50 Mbps
plan ($2.90/Mbps/month).

53. To estimate the effect of reduced prices on broadband penetration in these areas,
we rely on Atkinson, et al, who find that a $1/month reduction in price per megabit increases
broadband penetration by 2.4 percentage points. 20 We assume the full effect of reduced prices
would be felt over four years, beginning once the deployment has been made (i.e., at the end of
each year). Thus, a $3.67 reduction in price/Mbps would result in an 8.81 percentage point
increase in broadband penetration by the end of the third year of our projection period.21

54. The impact of increased availability of any type of broadband can be estimated
more directly. According to Morgan Stanley, the national residential broadband penetration rate
is approximately 56 percent of all households as of 2009, and is forecasted to increase to 61.1
percent as of 2011.22 We assume that households who receive broadband availability as a result
of the rural/underserved tax expensing proposal will begin subscribing to broadband in the year
following deployment, and that once subscriptions begin, they will subscribe to broadband at the
national average rate over the course of three years, i.e., that 20 percent of households will
subscribe in the first year, 40 percent in the second year, and 60 percent in the third year. Under
this assumption, 20 percent of all homes passed by broadband for the first time as a result of the
rural/underserved tax expensing proposal would be subscribers as of 2011. In addition, we
assume that 10 percent of all homes passed by fiber as a result of the various 100 /20 Mbps
proposals would be located in areas that would not have broadband availability without the
expanded fiber deployment. Therefore, we assume that 20 percent of those newly passed homes
become broadband subscribers by the end of 2011.

55. Finally, to estimate the impact of increased broadband penetration on


employment, we rely on the results of a 2007 study published by the Brookings Institution. In
that study, Robert Crandall, William Lehr and Robert Litan found that a one percentage point
increase in broadband population penetration (defined as broadband lines per person) will

20 Robert D. Atkinson, Daniel K. Correa, and Julie A. Hedlund, Explaining International Broadband
Leadership, Information Technology and Innovation Foundation (May 2008)
21 For example, if the number of households passed by FTTH increased by 1,000 as a result of one of the
proposals we examined, we estimate that 88.1 additional households become subscribers during the period of our
projection.
77 Simon Flannery, Benjamin Swinburne, David Gober, Daniel Gaviria, & Chad Harris, Morgan Stanley,
Cable/Sat & Telecom Broadband Outlook: Online Usage Growth Favors Cable, DirecTV Remains HD Leader (July
18, 2008), at Ex. 26.

17

increase private, nonfarm employment by 293,200 jobs (when the economy is not at full
employment).23

B. Results of Indirect Effects Analysis

56. Table 9 shows the results of our analysis of the effects of each proposal on
broadband adoption, and the resulting indirect economic effects on job creation. Specifically, we
find that the various proposals would increase the number of U.S. broadband subscribers by
between 268,800 and 3.39 million, increase the U.S. broadband penetration rate (defined as
broadband subscriber lines per person) by between 0.09 percent and 1.08 percent, and increase
employment by between 25,160 jobs and 317,000 jobs.

TABLE 9: SUMMARY OF INDIRECT EFFECTS ON JOBS CREATION, 2009-2011


50% Expensing for Private
511 Mbps Sector Tax- Public Sector
100% Expensing (Rural/Underserved Credit Tax Tax-credit
for 100/20 Mbps Areas only) Bonds Bonds
Additional Homes Passed (000)
- Fiber 2,995.1 - 6,552.5 - 34,114.2 4,523.5
- Any Broadband 299.5 - 655.3 1,343.9 - 3,952.7 3,411.4 452.4
Additional Subscribers (000)
- Fiber 297.3 - 650.5 - 3,386.6 449.1
- Any Broadband 297.3 - 650.5 268.8 - 790.5 3,386.6 449.1
Increase in Overall U.S.
Broadband Population 0.09% - 0.21% 0.09% - 0.25% 1.08% 0.14%
Penetration Rate24
Additional Jobs 27,831 - 60,888 25,160 - 73 ,999 317,000 42,034

VII. CONCLUSION

57. In this study, we have calculated the total economic impact of four different tax
incentive proposals relating to increasing broadband deployment and adoption. We find that each
of the four proposals generates substantial benefits to the U.S. economy through both increased
GDP and increased employment. Each of the tax proposals would directly result in thousands of
additional jobs sustained per year from 2009 to 2011. The number of new jobs sustained from
2009 to 2011 resulting directly from the private sector tax-credit bond proposal alone is as high

23 Robert Crandall, William Lehr, & Robert Litan, The Effects of Broadband Deployment on Output &
Employment: A Cross Sectional Analysis of US. Data, 6 ISSUES IN ECONOMIC POLICY 12-14 (July 2007).
24 Based on Morgan Stanley's baseline forecast of 75.156 million residential broadband subscribers in 2011
and the U.S. Census Bureau's U.S. population forecast of 313.2 million in 2011. Simon Flannery, Benjamin
Swinburne, David Gober, Daniel Gaviria, & Chad Harris, Morgan Stanley, Cable/Sat & Telecom Broadband
Outlook: Online Usage Growth Favors Cable, DirecTV Remains HD Leader (July 18, 2008), at Ex. 26; U.S. Census
Bureau, U.S. Population Projections, National Population Projections (Released 2008)
(http://www.census.gov/population/www/projections/downloadablefiles.html) .

18

as 197,437. These proposals result in even further job creation through their indirect effect of
increased broadband adoption. Given these proposals' relatively small impact on tax revenues
compared to the large resulting increases in GDP and employment, their long-run benefits in
increasing productivity and competitiveness, and their significant and virtually immediate impact
on economic activity,'' the adoption of any of these proposals would create substantial net
benefits to the U.S. economy.

'' For a discussion of the importance of timing in the effectiveness of fiscal stimulus policies, see Peter R.
Orszag, Options for Responding to Short-Term Economic Weakness, Testimony Before the Committee on Finance,
United States Senate (January 22, 2008), especially at 5 ("The timing of fiscal stimulus is critical. If the policies do
not generate additional spending when the economy is in a phase of very slow growth or a recession, they will
provide little help to the economy when it is needed.") and at 8 ("Tax cuts for business investment may be more
effective in boosting short-term demand if they are temporary than if they are permanent. Firms may view them as
one-time opportunities for tax savings, which may induce firms to move up some of their future investment plans to
the present.")

19